Source code for tradeexecutor.analysis.slippage

"""Display trading positions as Pandas notebook items."""
import binascii
import datetime
from _decimal import Decimal
from typing import Iterable

import numpy as np
import pandas as pd
from eth_abi import decode

from tradeexecutor.ethereum.revert import clean_revert_reason_message
from tradeexecutor.state.trade import TradeExecution
from tradeexecutor.utils.slippage import get_slippage_in_bps


def _ftime(v: datetime.datetime) -> str:
    """Format times"""
    if not v:
        return ""
    return v.strftime('%Y-%m-%d %H:%M')


def _decode_generic_adapter_execute_calls_args(data: bytes) -> dict:
    """Decode arguments for a generic adapter call.

    Needs to bulldoze through various warpped ABI encodings.
    """

    #   return encodeArgs(
    #     ['address[]', 'uint256[]', 'address[]', 'uint256[]', 'bytes'],
    #     [incomingAssets, minIncomingAssetAmounts, spendAssets, spendAssetAmounts, encodedExternalCallsData],
    #   );

    #all_args_encoded = encode(
    #    ["address[]", "uint256[]", "address[]", "uint256[]", "bytes"],
    #    [_addressify_collection(incoming_assets), min_incoming_asset_amounts, _addressify_collection(spend_assets), spend_asset_amounts, encoded_external_calls_data],
    #)

    integration_manager_abi = ["address", "bytes4", "bytes"]
    generic_adapter_address, selector, generic_adapter_payload = decode(integration_manager_abi, data)

    # incoming assets, min_incoming, spend assets, spendasset amounnts, encoded_external_calls_data
    generic_adapter_abi = ["address[]", "uint256[]", "address[]", "uint256[]", "bytes"]
    decoded_2 = decode(generic_adapter_abi, generic_adapter_payload)

    call_abi = ["address[]", "bytes[]"]
    called_addresses, called_data = decode(call_abi, decoded_2[4])

    calls = list(zip(called_addresses, called_data))

    return {
        "incoming_assets": decoded_2[0],
        "min_incoming": decoded_2[1],
        "spend_assets": decoded_2[2],
        "spend_assets_amounts": decoded_2[3],
        "encoded_external_calls_data": decoded_2[4],
        "calls": calls,
    }


[docs]def display_slippage(trades: Iterable[TradeExecution]) -> pd.DataFrame: """Format trade slippage details for Jupyter Notebook table output. Display in one table :return: DataFrame containing positions and trades, values as string formatted """ items = [] idx = [] t: TradeExecution for t in trades: idx.append(t.trade_id) flags = [] if t.is_failed(): flags.append("FAIL") if t.is_repaired(): flags.append("REP") if t.is_repair_trade(): flags.append("FIX") lag = t.get_execution_lag() reason = t.get_revert_reason() if reason: reason = clean_revert_reason_message(reason) tx_link = None block_number = None input = t.get_input_asset() output = t.get_output_asset() # Start with all columns NaN amount_in = np.NaN amount_out = np.NaN uniswap_price = np.NaN enzyme_expected_amount = np.NaN block_executed = np.NaN # Swap is always the last transaction if len(t.blockchain_transactions) > 0: swap_tx = t.blockchain_transactions[-1] block_number = swap_tx.block_number if swap_tx.function_selector == "callOnExtension": # Enzyme vault tx + underlying GenericAdapter wrapper # Assume Uniswap v3 always # wrapped args:[['2791bca1f2de4661ed88a30c99a7a9449aa841740001f47ceb23fd6bc0add59e62ac25578270cff1b9f619', '0x07f7eB451DfeeA0367965646660E85680800E352', 9223372036854775808, 3582781, 1896263219612875]] uni_arg_list = swap_tx.wrapped_args[0] uniswap_amount_in = uni_arg_list[-2] uniswap_amount_out = uni_arg_list[-1] # # Do a lot of decoding different parts of Enzyme transaction # and cross referencing different amounts # amount_in = input.convert_to_decimal(uniswap_amount_in) amount_out = output.convert_to_decimal(uniswap_amount_out) uniswap_price = amount_in / amount_out if t.is_sell(): uniswap_price = Decimal(1) / uniswap_price # TODO: Was legacy, is still around? # generic_adapter_data = binascii.unhexlify(swap_tx.args[2]) generic_adapter_data = swap_tx.transaction_args[2] enzyme_args = _decode_generic_adapter_execute_calls_args(generic_adapter_data) # Check we did not pass wrong token address to enzyme enzyme_incoming_token = enzyme_args["incoming_assets"][0] assert enzyme_incoming_token == output.address enzyme_min_incoming_raw = enzyme_args["min_incoming"][0] enzyme_expected_amount = output.convert_to_decimal(enzyme_min_incoming_raw) uniswap_function_selector = enzyme_args["calls"][0][1][0:4].hex() # exactInput((bytes,address,uint256,uint256,uint256)) assert uniswap_function_selector == "c04b8d59" block_executed = swap_tx.block_number tx_hash = swap_tx.tx_hash # TODO: Does not work in all notebook run times # tx_link = f"""<a href="https://polygonscan.io/tx/{tx_hash}>{tx_hash}</a>""" tx_link = tx_hash block_decided = t.price_structure.block_number or np.NaN items.append({ "Flags": ", ".join(flags), "Position": f"#{t.position_id}", "Time": t.executed_at or t.failed_at, "Trade": f"{input.token_symbol}->{output.token_symbol}", # "Started": _ftime(t.started_at), #"Executed": _ftime(t.executed_at), # "Block": f"{block_number:,}" if block_number else "", "Lag": lag.total_seconds() if lag else np.NaN, "Slippage tolerance": get_slippage_in_bps(t.slippage_tolerance) if t.slippage_tolerance else np.NaN, "amountIn": amount_in, "amountOut": amount_out, "Enzyme amountOut": enzyme_expected_amount, "Planned execution price": t.planned_price, "amountIn/amountOut price": uniswap_price, "Block decided": f"{block_decided:,}", "Block executed": f"{block_executed:,}", # "Notes": t.notes, "Failure reason": reason, "Tx": tx_link, }) df = pd.DataFrame(items, index=idx) df = df.fillna("") df = df.replace({pd.NaT: ""}) return df